1st International Workshop on

Machine Learning for Irregular Time Series (ML4ITS2021)

Co-located with Virtual ECML PKDD 2021

Submission HERE

About The Event

Call for papers

Time series data are ubiquitous. The broad diffusion and adoption of Internet of Things (IoT) and major advances in sensor technology are examples of why such data have become pervasive. These technologies have applications in several domains, such as healthcare, finance, meteorology, and transportation. Deep Neural Networks (DNNs) have recently been used to create models that improve on the state of the art for some of these tasks.

Time series data influences both political and industrial decisions every day, yet there is, surprisingly, limited research in Machine Learning (ML) for time series - especially in situations where data is scarce or of low quality. In many real-world applications, we have the following two scenarios: 1) the amount of available training data is limited, or 2) there is a huge amount of available data which is scarcely or not labeled due to high costs of data collection and annotation. As a result, the future of Artificial Intelligence (AI) will be about “doing more with less”. We believe there is a need for focusing on modern AI techniques that can extract value from such challenging situations. These considerations can also contribute to the increasing need to address sustainability and privacy aspects of ML and AI. In this context, heterogeneity of the data (e.g. non-stationarity, multi-resolution, irregular sampling) as well as noise, pose further challenges.

The main scope of this workshop is to advance the state-of-the-art in time series analysis for “irregular” time series. Here, we define time series to be “irregular” if they fall under one or several of the following categories:

  • Short: univariate and multivariate time series with a limited amount of data and history,
  • Multiresolution: multivariate time series where each signal has a different granularity or resolution in terms of sampling frequency,
  • Noisy: univariate/multivariate time series with some additional perturbation appearing in different forms. In this class, we also include time series with missing data,
  • Heterogeneous: multivariate time series, usually collected by many physical systems, that exhibit different types of embedded, statistical patterns and behaviours,
  • Scarcely labeled and unlabeled: univariate/multivariate time series where only a small part of the data is labeled or completely unlabeled.

This workshop intends to offer the ideal context for dissemination and cross-pollination of novel ideas in designing machine learning models suitable to deal with irregular time series. We specifically call for contributions addressing one (or more) of the irregularity aspects mentioned above. Accordingly, topics of interest for the workshop include, but are not limited to:

  • Methods for data imputation and Denoising,
  • Generative models for synthetic data generation
  • Transfer learning and transformer architectures for time series forecasting and classification,
  • Attention mechanisms for time series forecasting
  • Graph neural networks for anomaly detection and failure prediction,
  • Quantification of uncertainty,
  • Neural network architectures for time series classification and forecasting,
  • Unsupervised and self-supervised learning for different time Series related tasks,
  • Representation learning for time series,
  • Few-shot learning and time series classification in a low-data regime,
  • GANs for time series analysis (i.e. anomaly detection, data imputation, data augmentation, data generation, privacy),
  • Physical-informed deep neural networks for time series modeling,
  • (Deep) reservoir computing and spiking neural networks for time series and structured data analysis.


Program Chairs

  • Massimiliano Ruocco (SINTEF Digital / Norwegian University of Science and Technology)
  • Erlend Aune (BI / Norwegian University of Science and Technology)
  • Claudio Gallicchio (Univeristy of Pisa)

Program Committee

  • Sara Malacarne (Telenor Research)
  • Pierluigi Salvo Rossi (Norwegian University of Science and Technology)
  • Bjorn Magnus Mathisen (Sintef DIGITAL)
  • Per Gunnar Auran (Sintef DIGITAL)
  • Jo Eidsvik (Norwegian University of Science and Technology)
  • Leif Anders Thorsrud (BI)
  • Gard Spreeman (Telenor Research)
  • Pablo Ortiz (Telenor Research)
  • Vegard Larsen (BI / Norges Bank)
  • Stefano Nichele (Oslomet / Simula)
  • Filippo Maria Bianchi (UiT the Arctic University of Norway)
  • Juan-Pablo Ortega (St. Gallen University, Switzerland)
  • Azarakhsh Jalalvand (Ghent University-imec, Belgium and Princeton University, USA)
  • Benjamin Paaßen (Institute for Informatics, Humboldt-University of Berlin, Germany)
  • Petia Koprinkova-Hristova (Institute of Information and Communication Technologies, Bulgarian Academy of Sciences)


Papers must be written in English and formatted according to the Springer LNCS guidelines followed by the main conference. Regular and short papers presenting work completed or in progress are invited. Regular papers are expected to provide original and innovative contributions, should not exceed 16 pages including references. Short papers, describing innovative ongoing research showing relevant preliminary results, are maximum 6 pages. Papers must be submitted in PDF format online via an Easychair submission interface.

Each submission will be evaluated on the basis of relevance, significance of contribution and quality by at least three members of the program committee. Submitted papers cannot be identical, or substantially similar to versions that are currently under review at another conference, have been previously published, or have been accepted for publication. All accepted papers will be published at ceur-ws.org (indexed by e.g. google scholar) or within Springer LNCS proceedings depending on the number of submissions. Reviews are single-blind. At least one author of each accepted paper is required to attend the workshop to present.

The workshop’s schedule is tailored to an online event format. As such, we plan to integrate live sessions by video conferencing, pre-recorded videos and discussions through a communication platform such as Slack. The workshop will include 3 Invited Talks in video conferencing (30 minutes + 15 minutes Q&A). Authors of accepted “paper contributions” (both oral and posters) will be asked to send a 10 minute video to be made available through the communication platform. During Poster Sessions, attendees will be encouraged to initiate discussions around the presented contributions in dedicated communication platform channels, possibly engaging one-to-one meetings. Papers selected as contributed talks will be presented by a 15-minutes live video (available also offline) followed by a 5 minutes live Q&A (by video conferencing).

To further encourage the exchange of ideas and discussions, the workshop will conclude with a panel discussion including invited speakers and program committee members. We will collect questions via an online form from the audience. Lunch and coffee breaks will be coordinated with the official schedule of the conference. During coffee breaks, the attendees will be encouraged to socialize and engage discussions through the communication platform.


  • Paper Submission Deadline: July 16, 2021 (Postponed)
  • Acceptance notification: August 6, 2021 (Postponed)
  • Camera ready deadline: August 20, 2021 (Postponed)
  • Workshop program and proceedings online: September 1, 2021
  • Workshop Date: September 13, 2021

Event Speakers

Here are some of our speakers

Cesare Alippi

Cesare Alippi

Politecnico of Milano, Milano, Italy; Università della Svizzera italiana, Lugano, Switzerland

Boris Oreshkin

Boris Oreshkin

Unity Technologies

Event Schedule

Here is our event schedule

Introduction and opening remarks - Massimiliano Ruocco

Paper Presentation Chair: Claudio Gallicchio

- Homological Time Series Analysis of Sensor Signals from Power Plants - Luciano Melodia and Richard Lenz

- Adversarial Generation of Temporal Data: A Critique on Fidelity of Synthetic Data - Ankur Debnath, Nitish Gupta, Govind Waghmare, Hardik Wadhwa, Siddhartha Asthana and Ankur Arora

- Continuous-Discrete Recurrent Kalman Networks for Irregular Time Series - Mona Schirmer, Mazin Eltayeb and Maja Rudolph

Coffee Break and Online Socialization

Panel and concluding remarks - Erlend Aune

- Sara Malacarne - Research Scientist, Telenor Research, Norway

- Juan-Pablo Ortega - Research Scientist, St. Gallen University, Switzerland

- Are Haartveit - Intelecy, Norway

- Filippo Maria Bianchi - University of Tromsø, Norway

End of Workshop